JPH06244081A - Surface configuration detection method and projection exposure device - Google Patents

Abstract

PURPOSE:To improve detection accuracy of tilt, height, etc., of a sample surface by imaging reflection light from a sample surface at a position of conjugate with a sample surface and by detecting tilt, height, etc., of a sample surface by casting the imaged light on the sample surface again to guide its reflection light to a photodetector and by detecting tilt, height, etc., of the sample surface based on information of interference fringes on the photodetector. CONSTITUTION:A reticle 9 is illuminated by an exposure illumination system 81 and its circuit pattern is imaged on a surface of a wafer (detection object) 4 on a stage 7 by a reduction lens 8. A surface detection system 2 detects and processes surface configuration information of a wafer. Object light 16 is cast on the wafer 4 at an incidence angle of 88 deg. and its reflection light and reference beam 17 are reflected by a mirror 19, reflected again by a mirror 22 through lenses 20, 21 to make them move reversely in the original optical path and reflected by a beam splitter 14 in a direction of prisms 23, 23'. The prisms 23, 23' enlarge a beam interval and provide a specified angle simultaneously to inject beam to a lens 24. Two beams projected from the lens 24 cross each other at a position A which is conjugate with a reflection surface of the wafer 4 and interfere there.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and a projection exposure apparatus for detecting the surface shape of a flat substrate such as a semiconductor wafer, and more particularly to calculating the surface shape of a resist from optical interference fringe information by FFT. The present invention relates to a surface shape detection method for obtaining the surface inclination, height, etc., and a projection exposure apparatus for performing focusing by this method.

[0002]

2. Description of the Related Art In a pattern exposure process for miniaturized semiconductor integrated circuits, liquid crystal displays, etc., it is necessary to accurately detect the inclination and height of the surface of a wafer or substrate to accurately align the exposed surface. is there. For example, when a pattern with a line width of 0.5 μm is exposed by using the i-line of a mercury lamp in a semiconductor integrated circuit, the depth of focus is set to ± 1 μm or less in consideration of the warp and surface unevenness of the wafer. Slope of about 10μra
d, The height of the wafer surface must be 0.1 μm or less.

Japanese Patent Application No. 1-100026 discloses a laser.
There is disclosed a method of detecting the height, inclination, etc. of the resist surface from interference fringes obtained by obliquely injecting the S-polarized light of the light into the wafer surface and causing the reflected light to interfere with the reference light. That is, the inclination of the wafer surface is obtained from the pitch of the interference fringes, and the height of the wafer surface is obtained from the phase of the interference fringes.

With this method, the beam size of the laser light can be made extremely small, so that it is not necessary to increase the focusing angle, and since interference fringes are used, the spot diameter on the sensor due to the use of the condenser lens can be reduced. Since the light does not spread, the incident angle of incident light can be set to 85 ° or more, for example, 88 ° or the like to reduce the amount of light penetrating into the photoresist.

Further, the reflected light from the wafer is reflected by the mirror and re-incident on the wafer surface so as to enlarge the phase difference between the reflected light and the reference light, and the intensity distribution of the interference fringes is accurately measured. I was able to detect it.

[0006]

SUMMARY OF THE INVENTION However, the above-mentioned Japanese Patent Application No. 1
In the method disclosed in Japanese Patent Laid-Open No. 100026, the wafer surface and the interference fringe detection position need to be maintained in a conjugate positional relationship, but the mirror has a conjugate positional relationship with the interference fringe detection position at the second incidence. The problem is that the wafer surface and the interference fringe detection position cannot be maintained in a conjugate positional relationship. Therefore, the incident light of the second time is defocused on the surface of the wafer to generate an offset, and this offset needs to be corrected each time the measurement is performed.

FIG. 8 is a diagram for explaining the above conventional interference fringe detection method. FIG. 8A shows the case where the incident light is reflected once on the wafer surface, and FIG. 8B shows the case where the incident light is reflected twice.
It is assumed that the wafer surface 41 has a protrusion as shown in the drawing with respect to the reference surface 40. The coherent laser light is divided into two, and one of them is used as detection illumination light 16 for the wafer surface 4
1 is irradiated, and the other is made the reference light 17.

In FIG. 8A, the detection illumination light 16 is reflected by the wafer surface 41, and the reflected light 46 is the reference light 1 on the surface x.
7, the interference fringes as shown by the solid line in FIG. In contrast to this, when there is no protrusion, the detection illumination light 16 is reflected by the wafer surface 41 as shown by the dotted line, so that it overlaps with the reference light 17 at a position deviated from the above x and is shown by the dotted line in FIG. Such interference fringes are generated.

Since the phase difference φ between the solid line and the dotted line in FIG. 3C is derived as in the equation (1), the phase difference φ between the interference fringes from the reference plane 40 and the interference fringes from the wafer surface is measured. Thus, the height z (y) of the protrusion on the wafer surface 41 can be obtained. φ = 4πm · cos θ · z (y) / λ (1) Note that m is the number of reflections and is 1 in the case of FIG.
θ is the incident angle of the detection illumination light 16 and λ is the same wavelength.

When m is 1, the phase difference φ is too small and the measurement accuracy is low. Therefore, as shown in FIG.
Is returned to the original optical path by the mirror 40 and reflected again on the wafer surface, and the second reflected light is superposed on the reference light 17 on the surface x to generate interference fringes.
Since m becomes 2 and the phase difference φ can be doubled, the measurement accuracy can be doubled.

However, the position of the reflecting surface of the wafer surface 41 and the surface x
If the optical system can be constructed so as to be conjugate with each other, the heights of the respective reflecting surfaces of the wafer can be associated with each other based on the interference fringe information, and the tilt of the wafer can be accurately detected. Since the positional relationship is conjugate with the interference fringe detection position, the wafer surface and the surface x deviate from the conjugated positional relationship,
As a result, the height of the reflecting surface position cannot be detected accurately, which is a problem.

That is, as shown in FIG. 9, the phase of the interference fringe intensity corresponds to the abrupt change of the height of the wafer surface shown in FIG. As shown in (d), since the phase change becomes gentle as shown in (c) of the figure, the height of the wafer surface calculated from this is as shown in (d) of the figure. It was a gentle one. An object of the present invention is to provide a method for detecting the surface shape of a wafer that can solve the above problems, and a projection exposure apparatus that can control the height and inclination of the exposure apparatus based on this information.

[0013]

In order to solve the above problems, the light beam of a coherent monochromatic light source shaped into a predetermined shape is divided into two, and one of the light beams (object light) is In a surface shape detection method for detecting the inclination, height, etc. of the sample surface from interference fringe information obtained by superimposing the reflected light obtained by irradiating the sample surface with the other optical beam (reference light) ,
Light reflected from the sample surface is imaged at a position conjugate with the sample surface, the imaged light is irradiated again to the sample surface, and the reflected light is provided at a position conjugate with the sample. The inclination, height, etc. of the sample surface are detected from the interference fringe information obtained by superimposing the reference light on the photodetector and superimposing it on the photodetector.

Therefore, an imaging optical system for forming an image of the reflected light from the sample surface at a position conjugate with the sample surface, and a reflection for re-irradiating the image light of the imaging optical system on the sample surface. Means, a detection optical system for guiding the reflected light of the re-irradiated light on the sample surface to a photodetector provided at a position conjugate with the sample, and the reference light for the reference light reflected on the photodetector surface And a signal processing circuit for detecting the inclination, height, etc. of the sample surface from the interference fringe information detected by the photodetector, and the inclination of the sample surface by the output of the signal processing circuit. , The position of the sample surface is properly corrected by means of controlling the height and the like.

The imaging optical system is composed of at least first and second lenses having focal lengths f ₁ and f ₂ , respectively, and the first lens is arranged at a distance f ₁ from the sample surface. , The distance between the second lens and the first lens (f
₁ + f ₂ ) apart from each other, and further the above-mentioned reflecting means is
The lens is placed at a distance f ₂ from the lens.
Further, the imaging optical system is composed of a lens having a focal length of at least f ₁ , and the lens is separated from the sample surface by a distance f.
Only ₁ apart place, the reflecting means so as to spaced apart by a distance f ₁ from the lens. Further, means for adjusting the crossing angle of the object light and the reference light on the photodetector is provided.

[0016]

After the reflected light from the sample surface is imaged at a position conjugate with the sample surface, it is re-irradiated on the sample surface and the reflected light is guided to the photodetector at a position conjugate with the sample. To form an interference fringe pattern with no phase error. The photodetector detects this interference fringe pattern, and the signal processing circuit calculates the tilt, height, etc. of the sample surface from the output signal of the photodetector, and corrects the tilt, height, etc. of the sample surface. Further, the lens system of the image forming optical system forms an image of the reflected light from the sample surface at a position conjugate with the sample surface, and the reflecting means reflects this image formation to form the sample surface via the lens system. Image.

[0017]

1 is a block diagram of an embodiment of a reduction projection exposure apparatus according to the present invention. The reticle 9 is illuminated by the exposure illumination system 81, and its circuit pattern is imaged on the surface of the wafer (inspection object) 4 on the stage 7 by the reduction lens 8, and the surface detection system 2 detects the surface shape information of the wafer. And process. In the present invention, for example, the line width of the circuit pattern on the wafer 4 is 0.35.
In the case of μm, the depth of focus on the wafer is ± 1 μm or less, and the tilt and height of the wafer 4 are 5 μrad and 0.0, respectively.
Control to 5 μm. Although FIG. 1 shows only one axis of the wafer tilt detection, there is actually a system for detecting the wafer tilt in the direction perpendicular to the paper surface.

The light emitted from the laser (coherent light source) 1 passes through a shutter 5 and its s-polarized (linearly polarized) component is a polarized light beam.
The beam is extracted by the beam splitter 6, the beam diameter is enlarged by the lenses 10 and 12, and the beam is separated into two parallel beams by the prism 13. Further, the opening 11 is located at a position conjugate with the reflecting surface of the wafer 4, and its shape determines the spot shape of the wafer surface. The prism 15 gives a predetermined angle to each of the two parallel beams. One of the light emitted from the prism 15 is applied to the wafer 4 as the object light 16,
The other becomes the reference light 17.

The object light 16 is irradiated onto the wafer 4 at an incident angle of 88 ° (88 ° with respect to the vertical line standing on the wafer 4), and its reflected light and reference light 17 are reflected by the mirror 19 and the lens 2
It is reflected again by the mirror 22 through 0, 21 and travels backward in the original optical path, and the beam splitter 14 causes the prisms 23, 2 to travel.
It is reflected in the 3'direction. The prisms 23 and 23 'widen the interval between the beams and at the same time give them a predetermined angle to make them incident on the lens 24. Two beams emitted from the lens 24
The beams intersect with the reflecting surface of the wafer 4 at a conjugate position A and interfere there.

A diaphragm 25 is provided at the position A, and the light passing therethrough is compressed in one direction by a cylindrical lens 27 via a lens 26 to form an interference fringe on the CCD sensor 28 surface. In the conventional device, the lens 2
0, 21 and the mirror 22 are omitted, and the reflected light from the wafer 4 is reflected by the mirror 19 which has a different mounting angle in principle and is re-incident on the wafer 4, and the reflected light is similarly reflected at another angle. It was configured to interfere with the reference light incident from. However, in the present invention, since the image of the wafer 4 is formed on the mirror 22 and reflected, the mirror 22 can be set at a position conjugate with the reflecting surface of the wafer 4.

FIG. 2 shows the wafer 4 and the lenses 20, 21,
It is a figure showing the positional relationship of a mirror. Figure 2
As shown in (a), the focal lengths of the lenses 20 and 21 are f ₁ and f ₂ , respectively, the distance between the wafer 4 and the lens 20 is f ₁ , and the distance between the lens 20 and the lens 21 is (f ₁ + f
₂ ) Set the distance between the lens 21 and the mirror 22 to f ₂ .

FIG. 3B shows the case where the wafer 4 is tilted by the angle ψ. The object light 16 is reflected by the wafer 4 with an angle inclined by 2ψ as compared with the case of (a). This reflected light 162 is reflected light 1 in the case of (a) by the lens 20.
It becomes parallel to 61 and the lens 21 makes the mirror 22 (a).
The reflected light 163 is condensed at the same place as in the case of (1), becomes parallel to the light 161, and is reflected by the lens 20 at the same reflection position as the reflected light 162 on the wafer 4. As a result,
The reflection position of the wafer 4 with respect to the second reflected light 164 can be a conjugate position.

Since the reflected light 164 is inclined at an angle of 4ψ with respect to the incident light 16, the cross angle with the reference light 17 is also 4 °.
Only ψ changes. The change in the crossing angle can be obtained from the periodic change in the interference fringes, and the tilt angle ψ of the wafer 4 can be calculated from the change. FIG. 2C shows a case where the height of the wafer 4 changes by d. The reflected light 165 with respect to the object light 16 travels in parallel with the reflected light 161 of FIG.
The light is condensed at a position separated by 1 and reflected by the lens 21.
It becomes parallel to 1 again and is reflected by Mira-21. Since this reflected light travels back in the same optical path as 165 and is reflected by the wafer 4, the second reflected light 166 travels in the same optical path as the object light 16.

The optical path difference due to the height change d is 4 ds.
Since it becomes in θ, and the phase of the interference fringes changes accordingly, the height change d can be calculated from this to calculate the cross-sectional profile of the wafer. If the tilt angle is present in addition to the height change d, the tilt angle ψ is first corrected and then the height change d is derived. Figure 2
(D) is a development view of the imaging relationship in (a) to (c) above. Since the right wafer image is inverted on the mirror 22 by the lenses 20 and 21, it is equivalently formed as the left wafer image.

Further, the left wafer image reflected by the mirror 22 is inverted at the same position on the wafer 4 by the lenses 21 and 20, so that the second reflection image is the same as the first reflection image. Therefore, the wafer image can be detected by detecting the second reflection image.

In FIG. 1, the reference light 17 is applied to the second reflected light shown in FIG. 2 to form an interference fringe,
The CCD sensor 28 detects the waveform, and the signal processing circuit 3
Is converted into a digital signal by A / D conversion, and the height / tilt control unit 50 calculates the height, tilt, etc. of the chip from this surface shape information and controls the stage 7 to tilt the height of the wafer 4. So that it matches the focal plane. The shape of the object light 16 on the wafer surface is given by the opening 11.

FIG. 3 shows an example of the shape of the object beam 16 on the wafer surface. When the shape of the opening 11 is circular, oblique irradiation results in elliptic light 32 extending in the illumination direction on the wafer 4 as shown in FIG. When the chip 33 on the wafer 4 and the exposure area 34 are set as shown in the drawing, the elliptical light 32 is irradiated in a diagonal direction of the chip 33. In this case, a small circular beam facilitates adjustment of the optical system.

FIG. 3B shows the case where the opening 11 is an elongated rectangle. The wafer 4 is illuminated with rectangular light 35 extending in a direction perpendicular to the illumination direction. In this case, since the depth of focus can be reduced, the lateral resolution of the surface shape can be increased. FIG. 3C shows the case where the opening 11 is rectangular. This light becomes a rectangular light 36 spread on the wafer 4 and can be applied to the entire surface of the exposure area 34.
The surface shape of the chip 33 can be detected by the sensor 28.

FIG. 4 is a block diagram of another embodiment of the projection exposure apparatus according to the present invention. In FIG. 4, the lens 20 shown in FIG.
21, mirror 22, lens 29, corner cube 3
0, replaced with Mira-31. The object light 16 reflected by the wafer 4 is reflected by the mirror 19, and a wafer image is formed on the reflection surface of the corner cube 30 which is at a position conjugate with the reflection surface of the wafer 4 via the lens 29. To do. Corner Cure
The reflected light from the beam 30 is returned on the optical path, and is imaged again on the wafer at the reflection position of the wafer to be reflected again. The reference light 17 is reflected by the folding mirror 31 and travels backward in the original optical path.

FIG. 5 shows the wafer 4 and the lens 2 in FIG.
It is a figure showing the positional relationship of 9, corner cube 30, mirror 31. As shown in FIG. 5A, these set the distance between the reflecting surface of the wafer 4 and the prism 30 with respect to the lens 29 to the focal length f ₁ of the lens 29.
FIG. 5B shows a case where the angle of the wafer 4 is inclined by ψ. The light 162 reflected by the wafer 4 becomes a light 162 traveling in parallel with the reflected light 161 of FIG. This reflected light 163 also travels in parallel with the light 161, and is reflected by the lens 29 at the same reflection position as the object light 16 on the wafer 4. As a result, the reflection position of the wafer 4 with respect to the second reflected light 164 can be set to a conjugate position.

Further, since the reflected light 164 is inclined at an angle of 4ψ with respect to the incident light 16, the crossing angle with the reference light 17 is also 4 °.
Only ψ changes. The change in the crossing angle can be obtained from the periodic change in the interference fringes, and the tilt angle ψ of the wafer 4 can be calculated from the change. FIG. 5C shows the case where the height of the wafer 4 changes by d. The reflected light 165 with respect to the object light 16 travels in parallel with the reflected light 161 in FIG. 5A to form an image on the corner cube 31, and the reflected light 166 travels backward on the original optical path.

At this time, since the optical path difference caused by the height change d becomes 4 dsin θ, the phase of the interference fringe changes accordingly. From this, the change d in height can be calculated to calculate the cross-sectional profile of the wafer. In addition, when the tilt angle is present in addition to the height change d, first,
After the inclination angle ψ is corrected, the height change d is introduced.

FIG. 5D is a development view of the image forming relationship in the above (a) to (c). The wafer image on the right is the lens 29.
Thus, the image is inverted on the corner cube 31, so that the image is equivalently formed as the left wafer image. Further, the left wafer image is reflected at the same position on the wafer 4 by the lens 29, so that the second reflected image is the same as the first reflected image. Therefore, the wafer image can be detected by detecting the second reflection image.

FIG. 6 is an explanatory diagram of an algorithm for calculating the surface shape of the wafer. The interference fringe waveform shown in (a) is subjected to fast Fourier transform (FFT) to obtain the spectrum shown in (b). At this time, the spectrum near the frequency f = 0 corresponds to the DC component b (x) of the interference fringe waveform, and the spectra at the positions of f ₀ and −f ₀ have fluctuations φ (x) in the phase of the interference fringe and amplitude a ( x).

After moving the spectrum corresponding to f ₀ to the frequency origin as shown in (c), inverse FFT is applied to obtain the waveform of (d). The absolute value of p (x) is the amplitude a (x) of the interference fringe, and the phase is the fluctuation φ of the phase of the interference fringe.
Represents (x). Since this phase has an uncertainty of π unit and further has a tilt offset corresponding to (f ₀ −f ₁ ), this phase is first traced along x to connect the phases.

Next, since the above f ₀ is determined by the crossing angle between the reflected light and the reference light 17, the tilt offset is corrected to obtain the phase fluctuation φ (x) shown in (f). This φ
By substituting (x) into the equation (1), the cross-sectional profile z (y) of the wafer 4 can be obtained as shown in the equation (2). z (y) = λφ (x) / 4πm cos θ (2) The coordinate x on the CCD 28 corresponds to the coordinate y on the wafer.

Further, the fluctuation of the phase of the interference fringes caused by the wavefront aberration of the optical system is the φ of the reference sample which is optically flat in advance.
(X) Store the data and store it in the wafer φ
Correct it by subtracting from (x). As a result, the tilt correction amount at the time of measuring the wavefront aberration is the same as that at the time of measuring the wafer surface, so that the tilt correction process performed in (e) and (f) of FIG. Further, as shown in FIG. 7, the data of one cycle of the interference waveform s (x) is converted into a sine wave (a ₀ + a ₁ s
In ω ₀ t + a ₂ cos ω ₀ t) can be matched with the least squares method to calculate φ (x). The inclination and height of the wafer surface are obtained from the surface shape data calculated as described above, and the stage mechanism 7 is controlled so that the wafer surface coincides with the focal plane. In addition, focusing can be performed accurately and functionally by selecting the surface shape information.

[0038]

According to the present invention, it is possible to improve the accuracy of detecting inclination, unevenness, etc. of the surface of a semiconductor wafer in a projection exposure apparatus. As a result, the wafer stage can be controlled so that the wafer surface can be accurately contained within the focus margin of the projection lens, so that the depth of focus of the optical beam irradiating the wafer surface will be reduced due to future finer patterns. It can cope with shallowing. Further, by referring to the design data and the pattern density of the reticle, focusing on a fine area of the pattern to be exposed, or excluding an area where the surface shape detection value is unstable, the wafer tilt from the information of the peripheral portion, The height and the like can be obtained and corrected.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment of a projection exposure apparatus according to the present invention.

FIG. 2 is an explanatory diagram of an optical system in FIG.

FIG. 3 is a diagram illustrating a relationship between a beam shape of object light with which a wafer is irradiated and a spot shape on the wafer.

FIG. 4 is a configuration diagram of another embodiment of the projection exposure apparatus according to the present invention.

5 is an explanatory diagram of an optical system in FIG.

6 and 7 are explanatory views of a surface shape calculation method according to the present invention.

FIG. 8 is an explanatory diagram of an interference fringe detection method in a conventional device.

FIG. 9 is a waveform diagram illustrating a problem of interference fringes detected by a conventional device.

[Explanation of symbols]

1 ... Laser, 2 ... Surface detection system, 3 ... Signal processing circuit, 4 ...
Wafer, 5 ... Shutter, 6 ... Deflection beam splitter, 7
… Stage, 8… Reduction lens, 9… Reticle,… 11…
Apertures, 13, 15, 23 ... Prism, 14 ... Beam splitter, 16 ... Object light, 17 ... Reference light, 18, 19, 2
2, 31, 40 ... Miller, 20, 21, 29 ... Lens,
25 ... Aperture, 27 ... Cylindrical lens, 28 ... CC
D sensor, 30 ... Convex, 40 ... Reference plane, 41
... wafer surface, 50 ... height / tilt control section.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Minor Yoshida, 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Kanagawa Prefecture Production Technology Research Laboratory (72) Inventor Ken Fujii, 882, Ichige, Katsuta-shi, Ibaraki Shares Company Hitachi Ltd. Measuring Instruments Division

Claims (7)

[Claims] 1. An optical beam of a coherent monochromatic light source shaped into a predetermined shape is divided into two, and one of the optical beams (object light) is divided.
Reflected light obtained by irradiating the sample surface with the other light beam
In the surface shape detection method for detecting the tilt, height, etc. of the sample surface from the interference fringe information obtained by superimposing the sample (reference light), the reflected light from the sample surface is placed at a position conjugate with the sample surface. The sample surface is irradiated with the imaged light again, and the reflected light is guided to a photodetector provided at a position conjugate with the sample. A surface shape detection method characterized in that the inclination, height, etc. of the sample surface are detected from the interference fringe information obtained by superposition. 2. An optical beam of a coherent monochromatic light source shaped into a predetermined shape is divided into two, and one of the optical beams (object light) is divided.
Reflected light obtained by irradiating the sample surface with the other light beam
In the projection exposure apparatus, the tilt and height of the sample surface are detected from the interference fringe information obtained by superimposing the sample (reference light) to correct the tilt and height of the sample. An image forming optical system for forming an image of reflected light from the sample surface at a position conjugate with the sample surface, a reflection means for re-irradiating the image forming light of the image forming optical system onto the sample surface, A detection optical system for guiding the reflected light on the sample surface to a photodetector provided at a position conjugate with the sample; a reference optical system for superimposing the reference light on the reflected light guided to the photodetector surface; A signal processing circuit for detecting the inclination, height, etc. of the sample surface from the interference fringe information detected by the photodetector, and means for controlling the inclination, height, etc. of the sample surface by the output of the signal processing circuit were provided. A projection exposure apparatus characterized by the above. 3. The image forming optical system according to claim 2, wherein the image forming optical system is composed of at least first and second lenses having focal lengths f ₁ and f ₂ , respectively, and the first lens is located at a distance f ₁ from the sample surface. The second lens is spaced apart from the first lens by a distance (f ₁ + f ₂ ), and the reflecting means is further spaced from the second lens by a distance f ₂ . A projection exposure apparatus characterized by the above. 4. The image forming optical system according to claim 2, wherein the image forming optical system is composed of a lens having a focal length of f ₁ and the lens is arranged at a distance f ₁ from a sample surface, and the reflecting means is the lens. A projection exposure apparatus characterized in that the projection exposure apparatus is arranged at a distance of f ₁ from. 5. The projection exposure apparatus according to claim 2, further comprising means for adjusting a crossing angle of the object light and the reference light on the photodetector. 6. A projection exposure apparatus for projecting and exposing a circuit pattern formed on a mask onto a substrate by a projection lens, wherein a beam substantially parallel to the side of the substrate is provided between the projection lens and the substrate. Irradiating means for irradiating a desired shape on the top, and an imaging optical system for forming an image of the reflected object light from the substrate irradiated by the irradiating means at a position conjugate with the substrate surface,
A detector that receives an optical image formed by the image forming optical system and converts it into a signal, and detects the surface shape of the substrate based on the signal detected by the detector to determine the inclination or height of the substrate. A projection exposure apparatus comprising: control means for controlling the surface of the substrate to substantially match the image plane of the projection lens. 7. A projection exposure apparatus for projecting and exposing a circuit pattern formed on a mask onto a substrate by a projection lens, wherein a beam substantially parallel to the side of the projection lens and the substrate is provided on the substrate. It is irradiated with a desired shape on the top, the reflected object light from the irradiated substrate is imaged at a position conjugate with the surface of the substrate, and the formed optical image is received by a detector and converted into a signal. Then, the surface shape detection method is characterized in that the surface shape of the substrate is detected based on the signal and the inclination or height of the substrate is controlled so that the surface of the substrate is substantially aligned with the image plane of the projection lens.